The applicant proposes the construction of a table top infrared microspectrometer that is capable of obtaining infrared spectra from a lOO nm X lOO nm region of a biological specimen in an aqueous medium. This represents an improvement of a factor of 4,000 over what commercial infrared microscopes are capable of accomplishing with conventional light sources. It also represents an improvement of roughly 30 below the diffraction limit. At this resolution, it should be possible to characterize sub-cellular organelles without having to use any stains or fluorescent labels, relying on the intrinsic vibrational absorption spectrum of the constituent biomolecules. The new infrared microscope depends on the availability of new tunable infrared laser sources, allowing for new instruments that previously required accelerator-based lasers. In collaboration with investigators at Princeton University, Stanford University and Boston University, the investigators have recently developed the world's first scanning near field infrared microscope based on free electron laser (FEL) to image biological samples. This new microscope has demonstrated that it is possible to break the diffraction limit in the infrared part of the spectrum, and has been used to acquire infrared images of both semiconductor and biological tissue sections. Most promising is the ability to obtain, for the first time, infrared images under water. The investigators propose that a table-top version of the new microscope can be constructed that can result in vibrational spectra from a single location in a sample on length scales that are unprecedented. The proposal requires a confluence of novel developments in many fields, relying on breakthroughs in lasers, scanning probe microscopes, new materials, and the adaptation of all of this to biological samples. At the same time, the payoff of the new technique is also extremely promising, because of its reliance on the intrinsic spectral characteristics of biomolecules, and not on externally applied stains or perturbing labels.